Feat/typo one word

tab_err/error_type/_typo.py

@@ -135,6 +135,9 @@ def typo(input_text: str, typo_error_period: int = 10, layout: str = "ansi-qwert
         message = "typo_error_period smaller than 1 is invalid, as multiple errors per word are not supported."
         raise ValueError(message)
 
+    if input_text == "":  # return random char if empty string
+        return random.choice(list(neighbors.keys()))
+
     splits = input_text.split(" ")
 
     # draw only from splits that have a content

tests/api/test_high_level_api.py

@@ -0,0 +1,106 @@
+import numpy as np
+import pandas as pd
+import pytest
+
+from tab_err.api.high_level import create_errors
+
+
+class TestHighLevelAPI:
+    """Tests the high-level API."""
+
+    def test_create_errors_basic(self, test_data: dict[str, pd.DataFrame]) -> None:
+        """Test that create_errors returns two DataFrames with expected properties."""
+        seed = 42
+        error_rate = 0.5
+        modified_data_4rows_5columns, data_4rows_5columns_error_mask = create_errors(test_data["data_4rows_5columns"], error_rate, seed=seed)
+        modified_data_10rows_3columns, data_10rows_3columns_error_mask = create_errors(test_data["data_10rows_3columns"], error_rate, seed=seed)
+
+        # Check that they are still dataframes
+        assert isinstance(modified_data_4rows_5columns, pd.DataFrame)
+        assert isinstance(data_4rows_5columns_error_mask, pd.DataFrame)
+        assert isinstance(modified_data_10rows_3columns, pd.DataFrame)
+        assert isinstance(data_10rows_3columns_error_mask, pd.DataFrame)
+
+        # Check Shapes
+        assert modified_data_4rows_5columns.shape == test_data["data_4rows_5columns"].shape
+        assert data_4rows_5columns_error_mask.shape == test_data["data_4rows_5columns"].shape
+        assert modified_data_10rows_3columns.shape == test_data["data_10rows_3columns"].shape
+        assert data_10rows_3columns_error_mask.shape == test_data["data_10rows_3columns"].shape
+
+        # Assert the error masks contain only boolean values
+        assert data_4rows_5columns_error_mask.dtypes.apply(lambda dt: np.issubdtype(dt, np.bool_)).all()
+        assert data_10rows_3columns_error_mask.dtypes.apply(lambda dt: np.issubdtype(dt, np.bool_)).all()
+
+        # Assert that the error masks have the correct proportion of True to False
+        assert pytest.approx(error_rate) == data_4rows_5columns_error_mask.to_numpy().mean()
+        assert pytest.approx(error_rate) == data_10rows_3columns_error_mask.to_numpy().mean()
+
+    def test_create_errors_seed(self, test_data: dict[str, pd.DataFrame]) -> None:
+        """Test that create_errors returns the same dataframe when a seed is used."""
+        seed = 42
+        error_rate = 0.5
+
+        modified_data_1, error_mask_1 = create_errors(test_data["data_10rows_3columns"], error_rate=error_rate, seed=seed)
+        modified_data_2, error_mask_2 = create_errors(test_data["data_10rows_3columns"], error_rate=error_rate, seed=seed)
+
+        # Ensure same seed yields same dataframes
+        pd.testing.assert_frame_equal(modified_data_1, modified_data_2)
+        pd.testing.assert_frame_equal(error_mask_1, error_mask_2)
+
+    def test_create_errors_error_rates(self, test_data: dict[str, pd.DataFrame]) -> None:
+        """Test that create_errors returns two DataFrames with expected properties."""
+        seed = 42
+        for i in range(11):
+            error_rate = 0.1 * float(i)
+            modified_data_100rows_3columns, data_100rows_3columns_error_mask = create_errors(test_data["data_100rows_3columns"], error_rate, seed=seed)
+            modified_data_10rows_3columns, data_10rows_3columns_error_mask = create_errors(test_data["data_10rows_3columns"], error_rate, seed=seed)
+            modified_data_10rows_3columns_with_datetime, data_10rows_3columns_with_datetime_error_mask = create_errors(
+                test_data["data_10rows_3columns_with_datetime"], error_rate, seed=seed
+            )
+
+            # Assert that the error masks have the correct proportion of True to False
+            assert pytest.approx(error_rate) == data_100rows_3columns_error_mask.to_numpy().mean()
+            assert pytest.approx(error_rate) == data_10rows_3columns_error_mask.to_numpy().mean()
+            assert pytest.approx(error_rate) == data_10rows_3columns_with_datetime_error_mask.to_numpy().mean()
+
+    def test_create_errors_more_models(self, test_data: dict[str, pd.DataFrame]) -> None:
+        """Test that when more error models are introduced, the create_errors method has expected DataFrame return."""
+        error_rate = 1.0
+        seed = 42
+        n_error_models = 2
+
+        modified_data_4rows_5columns, data_4rows_5columns_error_mask = create_errors(
+            test_data["data_4rows_5columns"], error_rate, n_error_models_per_column=n_error_models, seed=seed
+        )
+        modified_data_10rows_3columns, data_10rows_3columns_error_mask = create_errors(
+            test_data["data_10rows_3columns"], error_rate, n_error_models_per_column=n_error_models, seed=seed
+        )
+        modified_data_100rows_3columns, data_100rows_3columns_error_mask = create_errors(
+            test_data["data_100rows_3columns"], error_rate, n_error_models_per_column=n_error_models, seed=seed
+        )
+
+        # Check that they are still dataframes
+        assert isinstance(modified_data_4rows_5columns, pd.DataFrame)
+        assert isinstance(data_4rows_5columns_error_mask, pd.DataFrame)
+        assert isinstance(modified_data_10rows_3columns, pd.DataFrame)
+        assert isinstance(data_10rows_3columns_error_mask, pd.DataFrame)
+        assert isinstance(modified_data_100rows_3columns, pd.DataFrame)
+        assert isinstance(data_100rows_3columns_error_mask, pd.DataFrame)
+
+        # Check Shapes
+        assert modified_data_4rows_5columns.shape == test_data["data_4rows_5columns"].shape
+        assert data_4rows_5columns_error_mask.shape == test_data["data_4rows_5columns"].shape
+        assert modified_data_10rows_3columns.shape == test_data["data_10rows_3columns"].shape
+        assert data_10rows_3columns_error_mask.shape == test_data["data_10rows_3columns"].shape
+        assert modified_data_100rows_3columns.shape == test_data["data_100rows_3columns"].shape
+        assert data_100rows_3columns_error_mask.shape == test_data["data_100rows_3columns"].shape
+
+        # Assert the error masks contain only boolean values
+        assert data_4rows_5columns_error_mask.dtypes.apply(lambda dt: np.issubdtype(dt, np.bool_)).all()
+        assert data_10rows_3columns_error_mask.dtypes.apply(lambda dt: np.issubdtype(dt, np.bool_)).all()
+        assert data_100rows_3columns_error_mask.dtypes.apply(lambda dt: np.issubdtype(dt, np.bool_)).all()
+
+        # Assert that the error masks have the correct proportion of True to False
+        assert pytest.approx(error_rate) == data_4rows_5columns_error_mask.to_numpy().mean()
+        assert pytest.approx(error_rate) == data_10rows_3columns_error_mask.to_numpy().mean()
+        assert pytest.approx(error_rate) == data_100rows_3columns_error_mask.to_numpy().mean()

tests/api/test_low_level_api.py

@@ -0,0 +1,24 @@
+import pandas as pd
+import pytest
+
+from tab_err import error_mechanism, error_type
+from tab_err.api.low_level import create_errors
+
+
+class TestLowLevelAPI:
+    """Tests the low-level API."""
+
+    def test_create_errors_error_rates(self, test_data: dict[str, pd.DataFrame]) -> None:
+        """Test that create_errors returns two DataFrames with expected properties."""
+        for i in range(11):
+            error_rate = 0.1 * float(i)
+            modified_data_100rows_3columns, data_100rows_3columns_error_mask = create_errors(
+                test_data["data_100rows_3columns"], "A", error_rate, error_mechanism.ECAR(), error_type.AddDelta()
+            )
+            modified_data_10rows_3columns, data_10rows_3columns_error_mask = create_errors(
+                test_data["data_10rows_3columns"], "A", error_rate, error_mechanism.ECAR(), error_type.AddDelta()
+            )
+
+            # Assert that the error masks have the correct proportion of True to False -- Note only one column is errored
+            assert pytest.approx(error_rate / 3.0) == data_100rows_3columns_error_mask.to_numpy().mean()
+            assert pytest.approx(error_rate / 3.0) == data_10rows_3columns_error_mask.to_numpy().mean()

tests/conftest.py

@@ -0,0 +1,19 @@
+import numpy as np
+import pandas as pd
+import pytest
+
+
+@pytest.fixture
+def test_data() -> dict[str, pd.DataFrame]:
+    """Fixture to provide test data before each test runs."""
+    rng = np.random.default_rng(42)
+    return {
+        "data_10rows_3columns": pd.DataFrame({"A": rng.integers(0, 100, 10), "B": rng.random(10), "C": rng.choice(["X", "Y", "Z"], 10)}),
+        "data_4rows_5columns": pd.DataFrame(
+            {"A": rng.integers(0, 100, 4), "B": rng.random(4), "C": rng.choice(["X", "Y", "Z"], 4), "D": rng.integers(0, 100, 4), "E": rng.random(4)}
+        ),
+        "data_100rows_3columns": pd.DataFrame({"A": rng.integers(0, 100, 100), "B": rng.random(100), "C": rng.choice(["X", "Y", "Z"], 100)}),
+        "data_10rows_3columns_with_datetime": pd.DataFrame(
+            {"A": rng.integers(0, 100, 10), "B": pd.date_range(start="2025-03-04", periods=10, freq="2h"), "C": rng.choice(["X", "Y", "Z"], 10)}
+        ),
+    }

tests/error_type/test_typo.py

@@ -0,0 +1,15 @@
+import pandas as pd
+
+from tab_err import error_mechanism, error_type
+from tab_err.api.low_level import create_errors
+
+
+def test_typo() -> None:
+    """Test that Typo replaces empty strings with a random character."""
+    test_data = pd.DataFrame(
+        {
+            "A": ["", "Alice", "Bob", "Bob", "Clara", "David"],
+        }
+    )
+    modified_df, _ = create_errors(test_data, "A", 1, error_mechanism.ECAR(), error_type.Typo())
+    assert modified_df.iloc[0, 0] != ""